To successfully complete the essential transition to a sustainable civilization, we must use both established and emerging technology. In this regard, the smart grid comprises multiple components and is closely linked to the notion of the Internet of Things, both of which are required for the transition by utilizing data to optimize system operations. The author's goal in this work is to lay out the current state-of-the-art of applications pertaining to energy management systems in buildings, including those incorporating IoT. After the initial discussion of the first chapter, a framework for the IoT is explored, followed by a discussion on smart buildings. The work continues with direct applications within buildings systems that are enhanced or enabled by the IoT, followed by a discussion on future trends and challenges. The work is wrapped up in a conclusion.
INTRODUCTION
Many issues of traditional energy systems are clear and, as we become aware of different problems concerning pollutants in the most adopted energy sources - which still account for over 78% of the world's energy consumption [1], a shift in this unsustainable scenario is necessary. In addition to that, many of today's standard applications and loads are, in fact, not efficient as they could be. Thus, a parcel of that generated and transported energy is, in fact, supplying a system which is inherently inefficient.
In addition to that, the human civilization will continue to demand more energy, increasing as years pass by. As life conditions improve in emerging regions, more energy is expected in its various sectors [2]. Additionally, as countries experience a growth in GDP, it is very likely that economic progress puts increasing pressure on energy use [3]. These factors are, currently, in a pattern leading to fossil fuel exhaustion and high emissions. To cope with that, it is required not only to migrate to a cleaner electricity generation matrix, but also improve efficiency in the operation of systems.
In the US alone, buildings account for over 40% of energy consumption, whereas the world's average is 24% - but rapidly raising due to population growth, enhancement on services sector, comfort level and increasing time spent indoors [2]. On average, energy usage in buildings accounts for 57% from to heating, ventilation and air conditioning (HVAC), and 22% from lightning systems, in the US [2]. Considering there is a great margin for efficiency improvement in these and other underlying systems in buildings, reductions in energy consumption are certainly reachable and advantageous, both by retrofitting existing buildings and including state-of-the-art practices and technology in new constructions [4].
Some of these emerging technologies are based on the concept of Internet of Things (IoT), which has been increasingly regarded on different topics, due to its ubiquitous sensing capacities and decision-making or system controlling based on real, abundant data.
As the IoT is consolidated, the necessity for understanding its potential applications by reviewing the current state-of-the-art exists, and will be herein investigated, focusing on Building Energy Management Systems (BEMS) usage. Previous reviews regarding IoT [5]-[9] focus on the network structure itself, its challenges and general literature, while [10]-[16] dwells on the IoT concept applied in smart homes and cities. Other papers discuss novel applications of individual BEMS systems, presenting details on HVAC [17]-[22], lightning [22]-[29], occupancy profiles [22], [30]-[35], demand response [36]-[45] and energy storage systems [46]-[48], and other services as water systems [49]. Additionally, an IoT framework has been designed and tested in smart energy for buildings [50]. However, no previous work has been done reviewing applications of this enabler IoT concept within a smart building environment.
The underlying potential of these applications range from individual settings, up to a wider application, as in a smart city composed by many of these buildings, offering required features of a renewable-intensive scenario - significant improve of efficiency, advanced demand response, resource forecasting, dynamic operation of resources, and energy storage [51].
Following this chapter, a framework for the IoT is explored in Chapter 2, investigating the concept as it is established today along with a literature review. Chapter 3 follows with a discussion on smart buildings, followed by investigating current state-of-the-art applications within BEMS, analyzing their potential, requirements, and efficiency and addressing their correlation with IoT as an enabler. Following up, a discussion on challenges is presented in Chapter 4, further discussing which problems are to be addressed in IoT applications at smart buildings. The work is wrapped up in a conclusion, on the last chapter.
IoT Framework
The trend of increasing processing power and decreasing size of electronic devices has been forecast to produce a number of sensors, widespread throughout all sorts of objects, rendering these as things, in which each smart object is uniquely addressable and has some capacity of communication and is connected to each other forming a network - the Internet of Things [16].
The term was first introduced in 1998, by Kevin Ashton, in a presentation at Procter & Gamble Company: "Adding radio-frequency identification and other sensors to everyday objects will create an Internet of Things, and lay the foundations of a new age of machine perception" [52]. In parallel, groups in Europe were using the term "ambient intelligence", later discontinued; all the while, in 2000 the term Internet of Things gained more popularity while being used by development groups of Auto-ID at MIT, in which Kevin Ashton participated as a founder [6], [7], [52].
Much attention has been drawn in recent years since then, evolving from its initial application on logistics [8]. The Council of the European Union, in 2008, has recognized and endorsed the work of the European Commission on the Internet of Things, inviting members and the Commission to explore challenges and opportunities within IoT emergence. Additionally, Cisco, Atmel, SICS and other leading technology vendors formed the IP for Smart Objects Alliance, announcing in 2008 the new uIPv6, a very small, open-source protocol stack, potentially enabling any device to possess an IP address, regardless of its power and memory limitations [7], [52].
More recently, consulting corporations have repeatedly extolled the Internet of Things as one of the most disruptive technologies by 2025, with profound impacts on U.S. economic development and military capability, as well as citizens' lives, business and global economy. SRI Consulting Business Intelligence, in 2008, has classified IoT as one among six disruptive technologies [53], and McKinsey Global Institute, in 2013, among twelve technologies [54].
In 2008-2009, according to the Cisco research groups, the number of internet-connected devices surpassed the number of people on Earth [55]. It is estimated the number of connected objects by 2025 may surpass 7 trillion; these include wearables and other things with which humans have direct interactions, but most will take place in infrastructural roles [6]. The existing architecture of Internet adopted in 1980, which uses TCP/IP protocols, cannot handle this number of devices while maintaining Quality of Service (QoS) [5].
This emphasizes the necessity of work and joint efforts to make the best uses out of this multitude of information available while shifting the traditional paradigm. The following sections discuss the two main characteristics of the IoT framework: Architecture, and Driver Technologies.
A. Driver Technologies
Most systems currently in use are managed in a vertical approach, with dedicated devices, infrastructure, and applications. The Internet itself is arranged in such way; it cannot handle the required information from so many devices as are predicted to take place. On the other hand, the IoT employs different standard protocols to achieve broadband networking while sharing infrastructure, physical and virtual environments. A layer commonly described as middleware is responsible for acting as an intermediate, handling the requests in between each other element [56]. In this section, an overview on key technologies is presented, structured in three main phases, followed by an investigation on the different architectures that have been proposed.
1) Collection phase: sensing of characteristics from the environment or objects takes place, collecting data within short-range in real time and processing it to be transmitted. Examples of sensing technologies are Radio-Frequency ID (RFID), infrared sensors, thermostats, GPS, cameras, and smart meters; sensors may be arranged in Sensor Networks. Short-range communication technologies may be open source standard, as WiFi, Bluetooth, ZigBee, M-BUS, Dash7, or proprietary solutions, as Z-Wave and ANT. The most infamous technology is the RFID, which was one of the precursors of IoT, and still is widely used due to its simplicity and efficacy [6].
A network composed of multiple wireless sensors is referred as Wireless Sensor Network (WSN), and is one of the key parts of IoT realization [9]. A finite number of sensitive nodes collecting data are mastered by a special purpose node making use of multi layered protocols organization [57]. These spatially distributed sensors which might range from hundreds or thousands, to millions, acquire ambient data as temperature, sound, vibration, pressure, motion, pollutants, illuminance, voltage, and others, providing the system with real-time information [7].
2) Transmission phase: the data is delivered to multiple external servers, throughout access of network. The available methods must then have access through heterogeneous gateways such as Ethernet, WiFi, xDLS, Cellular, Satellite, and PLC, and also make use of addressing and routing systems as LEACH, RPL, IP, and Trickle. The specifications of environment, physical limitations, speed and broadband requirements, previous existing infrastructure, in addition to the characteristics of each of these technologies, are to be accounted during design [6].
3) Processing, managing and utilization phase: this phase possesses processing power for analyzing the data and organizing the information flow, distributes it for different applications and services, and provides feedback for all control systems within the process. Other vital functions are also part of this stage, as device management and discovery, filtering, utilization and aggregation of data. Cloud computing, SOA, and P2P are enablers of this phase, and applications have a vast array of possibilities [6].
B. Architecture
With such complexity and ranging throughout many different areas, the IoT inherently has a problem of not having unified architectures [7]. Therefore, there is not one, but many different specifications on what is its architecture. A review of 129 research papers by [9] has shown the trend of architecture organization of IoT in different areas.
These architectures generally consist of the basic proposal of arrangement and communication between sensors, one or more networks, and computing technologies, set in a way so to achieve one or, generally, multiple goals, which are defined as applications or services.
For instance, three- [58], four- [59], five- [60] and six-layer [5], [61] architectures have been proposed. Other arrangements have also been proposed besides multilayered ones [7]. Additionally, application-specific architectures have been proposed, e.g. in healthcare [62], meter reading [63], enterprise services [64], agriculture and cloud service and management [9].
The excessive solutions are certain to difficult interoperability, resulting in slowing down of IoT development process [6]. To avoid further growing complexity, major standard bodies (3GPP, IEEE, ETSI) have worked on defining two major architectures.
1) Hierarchical network architecture for scalable connectivity: promoted by the ETSI Technical Committee for Machine-to-Machine communications (M2M), established in 2009 to develop and maintain and end-to-end (e2e) reference architecture for M2M networks. Along with scalability, it is designed on an IP-based architecture that is easily developed, simple, efficient, interoperable, makes use of standard protocols, masks underlying networks' complexity for applications developers, and promotes development of new services [65]. In its high-level architecture, the network is divided into three domains:
Device and Gateway Domain: composed of a high quantity of M2M devices interconnected in between themselves, one or more area networks and one or more gateways. Applications are executed by the devices, including sensing capabilities, which are connected within the area network through short-range technologies previously described. The gateway functions as a proxy between M2M devices and the network, and might locally execute applications on itself, gathering data and processing it.
Network Domain: formed by Access Networks and Core Networks, offering connectivity between the two other domains (M2M gateways and applications). Access Networks include xDSL, HFC, PLC, W-LAN, WiMAX, satellite, and UWB technologies; Core Networks provide additional features and include 3GPP GPRS and EPC.
Application Domain: consists of applications, either on the client or on the server side, middleware, business layer, plus enabler features that facilitate interactions between applications and devices.
2) Hierarchical network architecture for high capacity: designed to cope with high traffic of data, from a number of devices and with high amount of data interexchange (e.g. audio, video). This architecture seeks to support Machine-Type Communication (MTC), enveloping M2M interactions as well as interactions between humans and machines. Besides optimizing Long Term Evolution (LTE) technology for high capacity, this architecture researched by 3GPP also aims to lower operational costs of operators, reduce impact and complexity of handling large MTC groups, optimize network operations to mitigate battery usage on devices, and foster development of new MTC applications [66], [67].
This architecture is formed by three basic elements: MTC Devices, MTC Internetworking Function, and MTC Server. In between the internetworking and devices elements, each Operator Network performs the connection. Although similar to the previously described architecture, the high capacity relies on 3GPP to provide multi-tier connectivity with high mobility [6].
Different authors have discussed, also, new concepts from the IoT topology. The concept of Io (in which '*' refers to 'all' in computing) has been discussed in [9], where the author presents this hypothetical concept that would suit any architecture and combine both untouched and cumulated studied areas. It would also account for future technologies, e.g. the miniaturization of circuits, different material types, and Operational Systems (OS) specifically designed for IoT incursion.
Another work has recently introduced the Network of Things (NoT) concept [68], in which objects can occur in the physical or virtual space. The primitives of a NoT are divided into Sensor, Aggregator, Communicator, External Unility, and Decision Trigger. The author details the conditions for each of these primitives in details regarding the architecture of a NoT in its work. Also, the IoT is classified as a type of a NoT [68], thus sharing the same elements.
IoT Applications and Services within Smart Buildings
The term smart refers not to the equipment installed on a building, but also extends to what is done with such information - sensing the environment, analyzing behavior and adequately performing the control which would otherwise be done by humans [69]. IBM estimates that intelligent buildings may reduce water usage from 30% to 50%, and cut off CO2 emissions and energy consumption by 50% to 70% [49]. The two main areas in which IoT will bring an array of benefits and enable smart buildings, can be classified between services and energy management.
In the services domain, most or all implementations for the majority of buildings are arranged in a vertical manner, as described in section II, in which each specific service has a specific system (devices, infrastructure). Not only that, but once installed, such system is seldom updated or maintained, and is prone to failures, malfunctioning, and low efficiency. Examples are emergency systems and alarms, security and access control, and water management. Some other services are, too, related to energy management, such as lightning, elevators, and HVAC, which have a direct impact on the building's efficiency.
On the energy management side, besides the aforementioned loads, we include controllable plug loads (e.g. computers), energy storage, demand response, distributed generation, climate control and peak shaving. Also, these are all requirements for a larger scenario in which renewables take over higher share on the generation matrix, bypassing the inherently intermittent characteristics of renewable trough advanced control of resources [1], [14]. Since buildings are responsible for over 40% of US energy consumption [2], turning buildings smart would provide an essential resource for the smart grid to achieve its requirements [11].
As an enabler for implementation of a BEMS, IoT allows each of these solutions to be feasible, and simplifies the design and operation of applications [10]. The other potential underlying within IoT is the sharing of infrastructure as depicted on Fig. 1, reducing costs, and maximizing QoS, lifetime, and other benefits [13], bringing its potential to the attention of stakeholders and governmental entities alike. In addition to that, smart buildings are one of the pillars for progression towards a smart grid and smart city ambient [15], and have been gaining attention from an environmental point of view as such - it is envisioned that improving the US grid by 5% would be equivalent to cutting CO2 emissions of 53 million cars [70], so any work done there has a huge benefic potential.
Many application-specific control techniques for BEMS intrinsic systems have been explored in the literature, and others successfully implemented. The next sections will dwell on recent progresses, analyzing their potential within an IoT perspective applied into smart buildings.
A. HVAC
The most energy-consuming load in buildings is, undoubtedly, HVAC systems. In the U.S., it represents over 57% and in average 48% for European countries [2]. It is, therefore, an important target for enhancement in BEMS.
Active and passive thermal storage in buildings has been found to substantially decrease energy consumption especially during peak-hours in which the cost of energy is higher, even when imperfect weather and buildings models are provided to controllers [19]-[21].
Control techniques using occupancy information show that even simple actuations, as turning ventilation off in a room that is vacated, brings significant reduction on energy consumption, over 15% on savings. Using perfect occupancy profiles is moderately more beneficial [22].
Occupancy and temperature are the baseline for these control techniques, which are contemplated in modern sensors [71]. In addition to these parameters, other characteristics and weather forecast can also be obtained and the conjunct, processed by the end application in the architecture, then achieves multiple goals, boosting efficiency while maintaining QoS. While the control strategy of individual solutions are implementable individually, many benefits arise from using an IoT structure [16].
Energy efficiency is a major concern in BEMS. To this end, manufacturers of HVAC equipments have made deliberate efforts to increase the energy efficiency of their products [2]. Initially, the need for energy efficient HVAC appliances was informed by a rise in energy costs. However, today, the major concern is not only a reduction in energy costs, but also issues to do with environmental conservation [9]. It is also noted that the wellbeing and productivity of people living in houses with improved HVAC systems is enhanced. Governments have come up with regulations to ensure that buildings are installed with energy efficient HVAC equipment. A case in point is the EPA in the US, which has introduced stringent regulations over the years. To meet these requirements, BEMS are using hi-tech heating and cooling systems. For instance, the adoption of forced air systems is associated with energy savings of up to 17.5% [70].
B. Lightning
Lightning loads are the second most energy consuming loads in buildings [2], [27]. Although intensive efforts have been given by the U.S. Department of Energy to promote efficient strategies on lightning, such as occupancy sensors, since 2002 [72], there is only so much implemented, and the widespread sensors are not always up to par [28]. Many times, human behavior in conjunction with badly implemented strategies may reduce potential savings or ever result in higher energy consumption, e.g. when people fail to turn off light switches after installation of presence sensors [26], [27], [29]. In other implementations, the time delay is not well adjusted and results in low savings [27]-[29]. To overcome these downsides, a few strategies are applicable.
Scheduling of activities is one of them, and to perform this control, the system considers which activity is programmed at a given ambient, according to the time and the day, and turns the lights on or off according to that. In a more advanced manner, these controls might be complemented by occupancy sensing [22], or replaced by it [29]. It is important to note that manual control ensures proper functioning, and at any IoT application in smart buildings the user feedback is deterrent for the system settings.
The quality of occupancy sensing is a key to improve lightning control systems. Previous works have shown great potential of reduction as quality of sensors increase [22], and it has been established that more extensive sensing and data analysis - e.g. a sensor network instead of a single sensor per room - is necessary to boost the benefits of the control implementations [27].
Similarly, it has been shown that for daylight-linked control systems, which dim artificial lightning based on daylight presence and intensity, better results are shown using distributed sensor network. Using a single sensor per room provided either less reduction of energy or lower performance of the system, or both; whereas one sensor per fixture renders the best efficiency and most reliable QoS [23].
Furthermore, it has been recognized that individual lightning control applications perform better when combined. That can be as simple as having two kinds of sensors as inputs for a control scheme [27], or two or more concomitant control strategies [29], i.e. occupancy and daylight or daylight and scheduling.
Overall, the extensive monitoring of occupancy and ambient light - again, contemplated on modern sensors - are certain to provide all the data necessary for different applications to be successfully deployed. The literature agrees on improved efficiency when combining different smart lightning techniques, and an IoT structure would allow so inherently.
Moreover, most smart lightning implementations have shown a need for readjustment in parameters, sensor variables, and applications' algorithms [23], [27]-[29]. When systems required further adjustments, implementation costs rose, and in some cases the savings result was far from the desired due to imprecision in the implementation system. Therefore, a full-scale IoT implementation is undoubtfully beneficial, as it allows visualization of data and its processing in an application level for the building manager or application developer. The adjustment of system at initial deployment, as well as throughout its utilization, is inherent in the system architecture, maximizing the flexibility and the results on energy savings.
Energy savings can also be realized through the use of energy efficient light bulbs [29]. A case in point is the Federal Green Challenge program. The project is a component of EPA's Sustainable Materials Management Program [29]. A number of federal agencies have realized savings in energy consumption by using energy saving bulbs. For instance, within 12 months, the Lovell Federal Health Care Center achieved a reduction in its energy consumption by 15 percent. The management of the institution had replaced energy consuming fluorescent bulbs with light emitting diodes. The latter are more efficient as far as energy consumption is concerned. It is also important to note that BEMS can reduce energy consumption by encouraging the use of daylight [27], [29]. The strategy involves the use of skylights in buildings. It may also entail the use of strategically positioned windows. When this happens, the need to switch on lights during the day is reduced.
C. Ancillary Services
Frequency regulation and voltage control are some of the ancillary services whose control is deemed to shift from only central regulators, giving space for distributed prosumers to perform some of the necessary regulations (producers-consumers) [73]. By having a predictable load and behavior, buildings are prone to contribute to these services. and previous work have shown good potential from prosumers. HVAC loads can be used as frequency regulators [17], and both HVAC [18] and lightning systems [26] can contribute to voltage regulation by providing demand response. Other works have presented demand response capabilities from plug loads, based on the policy of building managers and users' restrictions; in one, a ZigBee-based network was established [44], and in another, Energy Information Gateways were used [45], while both used metering on individual plugs.
Solutions making use of OpenADR [43] and novel algorithms [36], as well as companies such as Comverge [37], [41] strongly encourage integration using IoT to provide demand response, in addition to show promising results in performing the demand response task. Since the information needed for control systems is dependent on a multitude of sensors and actuators, it is only advantageous to apply the IoT framework to group applications and maximize their capabilities.
Other opportunities for ancillary services arise from Plug-in Electric Vehicles (PEVs) and Electrical Energy Storage Systems (EESS). Although not necessarily always present in buildings, they are growing in presence and offer capabilities at peak shifting and voltage regulators, even more so in presence of high renewable penetration [46], [47]. Offering yet more of the required flexibility for a future smart grid, these systems are easily integrated inside the IoT structure - their information is already collected and managed locally, and further processing this data for additional purposes is intrinsic within its architecture [52].
Voltage stability is an important attribute of ancillary services as far as BEMS is concerned [52]. In this case, manufacturers of electrical equipments ensure that the devices can only function optimally at a given voltage value. A few of the electrical devices can accommodate slight fluctuations in the specified voltage [47]. However, significant fluctuations have negative impacts on the performance and life of the device. To achieve the goal of these ancillary services with regards to energy saving, some buildings are fitted with voltage regulators. However, it is the role of the power company to provide a stable output that meets the specified voltage levels. Frequency stability is another important component of ancillary services [46]. In this case, governments and other regulatory agencies require power grids to operate within specified frequencies. The range of frequency in most instances is set at 50Hz or 60Hz [46].
D. Other Services
Water management systems can vastly benefit from the IoT potentials [49]. For instance, it has been proposed the identification of leakage by placing sensors in strategic points of the system [12], which is normally hindered by inefficient detection methods thus entailing in water wastage. The same is true for gas leakages, or fire, at which point a smart building can identify through self-diagnostic and warns inhabitants and the proper authorities when relevant [69].
Healthcare services also have a big appeal presently, and IoT provides the possibility of integration that is necessary to promote all assistance and attention required for denizens in an individual level [62].
Additionally, all data acquired from sensor arrays may be used for novel applications, rendering new usage for this information and promoting yet more tools for building managers to efficiently have data-oriented decision making. Suggestions include indoor positioning, building information services, and building usage and energy maps [29]. The same information can also be used for security and access control purposes, while sharing some of the infrastructure.
Agriculture is another field that can benefit from the adoption of IoT [29], [69]. The adoption of this technology has led to the emergence of innovative farming techniques. The agricultural sector is facing a number of challenges. They include climate change and increased population. The challenges can be addressed through the adoption of IoT. One strategy entails linking wireless sensors to agricultural mobile apps and cloud platforms. The move helps in the collection of data touching on climate and such other elements that affect productivity. The elements include pest infestation and soil fertility. Furthermore, IoT can be used in environmental monitoring and conservation. For instance, IoT applications can be used to monitor the quality of air and water [62]. Environment conservationists can use the applications to monitor the migration and habitat of wildlife [12]. Finally, IoT applications can also be used as early-warning systems for such calamities as floods and earthquakes.
Challenges and Future Direction
Still a new topic, IoT presents an array of complex problems while fostering new developments for many areas.
Standardization, as discussed in Chapter 2, is a pressing concern which regulation agencies are aware of, and work is being developed to circumvent further multiplication of base concepts.
The rise of technologies as Micro-Electro-Mechanical Systems (MEMS) and continuous development of Nano, Optical, and ICT technologies [5], [9] continuously brings more complexity in modern society, altogether with their benefits. Unless their deployment and use is simplified and integrated - for one, through means of IoT concept - this array of technologies and advances will bring exponential increase in complexity of use and integration, hindering progress and potentially diminishing adoption of IoT systems.
The Europe Information Society Technologies Programme Advisory Group (ISTAG) has evidenced 5 research challenges, describing potential IoT success based upon how quickly and effectively investors overcome these. They are: Edge Technologies (sensors, actuators, tags, embedded systems), Networking Technologies, Middleware Systems (effective use real-world data in diverse applications), Platform Services (guarantees scalability, availability, security), and Web Service Technologies (providing information and services available, reducing interoperability issues and promoting platform independence) [52].
Another key aspect that must be addressed is security, both the protection of data and the privacy of users [5]. As much more information is going to be in transit, it is a necessity that every user has guarantee of privacy, and that all data is protected from exploits which might target users, applications, companies, or nations - which makes IoT to be regarded as a national security priority.
On top of those issues, the reliability of every system must be enforced, while also having a user-friendly approach and deliver the best QoS for the maximum number of users [16].
Regarding smart buildings, the increasing capacity and decreasing cost of sensors is transforming the way stakeholders and companies envisage building management. The "smartification" of things is already happening, and smart buildings are rising as IoT becomes relevant and feasible by research and technological advancements. Since studied systems ...
